• Aucun résultat trouvé

Reason-based choice and context-dependence: An explanatory framework

N/A
N/A
Protected

Academic year: 2022

Partager "Reason-based choice and context-dependence: An explanatory framework"

Copied!
59
0
0

Texte intégral

(1)

Reason-based choice and context-dependence:

An explanatory framework

Franz Dietrich & Christian List This version: 26 July 2015

Abstract

We introduce a “reason-based” framework for explaining and predicting individual choices. The key idea is that a decision-maker focuses on some but not all proper- ties of the options and chooses an option whose “motivationally salient” properties he/she most prefers. Reason-based explanations can capture two kinds of context- dependent choice: (i) the motivationally salient properties may vary across choice contexts, and (ii) they may include “context-related” properties, not just “intrinsic”

properties of the options. Our framework allows us to explain boundedly rational and sophisticated choice behaviour. Since properties can be recombined in new ways, it also o↵ers resources for predicting choices in unobserved contexts.

Keywords: Rational choice, reasons, context-dependence, bounded and sophisticated rationality, prediction of choice.

1 Introduction

How can we explain an agent’s choices? The classical theory of rational choice does so by ascribing to the agent a preference relation over the options – in the simplest case, an ordering. This preference relation explains the agent’s choices if, in every choice context, the agent chooses the most preferred option among the feasible ones.1 The choices are then said to berationalized by the preference relation. When choices involve uncertainty, we must ascribe beliefs as well as preferences to the agent, such that the agent always

Contact details: F. Dietrich, Paris School of Economics & CNRS, CES-Centre d’Economie de la Sorbonne, Maison des Sciences Economiques, 106-112 Boulevard de l’Hˆopital, 75647 Paris cedex 13, France; URL:<http://www.franzdietrich.net>. C. List, London School of Economics, Departments of Government and Philosophy, London WC2A 2AE, U.K.; URL:<http://personal.lse.ac.uk/LIST>.

1Or, if there is no unique most preferred option, he or she chooses one that is tied for most preferred.

To appear in Economics and Philosophy, 2016

(2)

chooses an expectation-maximizing option, but the logic of the explanation is similar.

Though elegant and influential, this theory has some well-known problems:

An empirical problem: It cannot accommodate all empirically documented patterns of choices. As psychologists and behavioural economists have amply shown, people often choose in ways that cannot be naturally rationalized by any preference relation over the options. For example, people are susceptible to framing e↵ects, often satisfice rather than optimize, and follow social norms that are not in line with the constraints of classical rational choice theory (e.g., Camerer et al. 2004). We give some illustrations later.

An explanatory problem: Even when there is a preference relation over the op- tions that rationalizes an agent’s choices, it is far from clear whether this can be viewed as a genuine explanation of those choices. For a start, many economists adopt a be- haviouristic interpretation of preferences and treat preference relations merely asformal representations of choices and not as genuinely explanatory. But aside from this concern, when we are asked, “why did you choose teaching rather than banking as your career”, simply saying “because I preferred one to the other” is not very illuminating. We are expected to give reasons for our choices, as philosophers and psychologists have long emphasized (e.g., Shafiret al. 1993; Lenman 2011). A better explanation might be that we perceive teaching as a way of making a social contribution and promoting learning, while we perceive banking as a way of making money and supporting the economy’s status quo; and we rank the first bundle of properties more highly than the second.

A predictive problem: A less widely recognized problem is that the classical theory is limited in its ability to predict an agent’s future choices (Bermudez 2009). If we simply ascribe a preference relation to the agent, based on his or her past choices, then we can predict future choices only in special cases: namely when this preference relation already ranks the options involved. This is only the case when these options are ones the agent has encountered before, unless we can somehow extrapolate the agent’s preferences to them. When the options are genuinely new, this extrapolation is difficult. This limitation is a byproduct of the parsimonious informational basis of classical choice theory.

We introduce a “reason-based” framework for explaining individual choices, which is intended to overcome all of these problems. It is prompted by our diagnosis of a key shortcoming of the classical theory: the lack of an account of how agents perceive the options they are faced with. In the classical theory, options are usually primitives, which are not further unpacked, and agents have preferences over them. In reality, however,

(3)

each option has numerous properties, and an agent focuses only on some, but not all, of these properties in making his or her choices. Recall the example of teaching versus bank- ing. An agent might perceive the first option as the property bundle “contributing to society and promoting learning” and the second as the property bundle “making money and supporting the economy’s status quo”. Our framework captures the idea that agents perceive the options in terms of “motivationally salient” properties. Choices are then made, not based on fixed preferences over options, but based on more fundamental pref- erences over motivationally salient property bundles (cf. Lancaster 1966, Gorman 1980).

We lift two common but problematic assumptions. One is that the agents whose choices we seek to explain perceive the options in the same way as we, the modellers, do. In our framework, we can express di↵erent hypotheses about how an agent perceives the options, and ask what choice behaviours these hypotheses would predict. A second assumption which we lift is that an agent will always perceive the same options in the same way, irrespective of the choice context. In our framework, an agent’s perception of the options may depend on the context, in the following two ways.

First, the motivationally salient properties may vary from context to context. We call this phenomenon “context-variance”. It arguably plays a role in framing e↵ects.

Second, the motivationally salient properties may go beyond “intrinsic” properties of the options and include “context-related” properties. Examples are whether an option conforms to a context-specific social norm (e.g., is it polite?), whether it is above average quality among the available options, or whether the choice menu o↵ers luxury options.

We call this phenomenon “context-relatedness”. It arguably plays a role in sophisticated choice behaviours such as non-consequentialist or norm-following behaviours.

Once we recognize those two kinds of context-dependence, we can explain many non- classical choice behaviours. Finally, the move from options as primitives to options that are perceived as bundles of properties also yields new resources for predicting an agent’s future choices: properties can be recombined in new ways, and an agent’s attitudes towards certain property instantiations in the past can give us evidence for his or her attitudes towards new instantiations of those properties.

Related literature: This paper is related to the large body of work on classical and non-classical choice theory in economics, psychology, and philosophy. For an overview of classical choice theory and the rationalization of choices by preferences, see Bossert and Suzumura (2010). There are, by now, many papers which propose non-classical models of individual choice, prompted by the shortcomings of standard rational choice theory (see, e.g., Sen 1993; Suzumura and Xu 2001; Kalai et al. 2002; Gaertner and

(4)

Xu 2004; Manzini and Mariotti 2007, 2012; Mandler et al. 2012; and Cherepanovet al.

2013). However, these works do not explain choices in the “reason-based” way developed here or in terms of the two orthogonal kinds of context-dependence we identify.2 There are some works discussing variants of one of those two kinds of context-dependence, notably papers by Salant and Rubinstein (2008), Bernheim and Rangel (2009), Bossert and Suzumura (2009), and Bhattacharyya, Pattanaik, and Xu (2011), as reviewed later.

An important precursor to our approach is Shafir, Simonson, and Tversky’s work on reason-based choice and context-dependent preferences in psychology (e.g., Simonson 1989, Shafir et al. 1993, Tversky and Simonson 1993; for a recent discussion, see de Clippel and Eliaz 2012). They proposed that “when faced with the need to choose, decision makers often seek and construct reasons in order to resolve the conflict and justify their choice” (Shafiret al. 1993: 11). Our framework can be viewed as a novel formalization and development of these ideas.

There are also several related works on property-based preferences, the logic of prefer- ences, and preference change. In consumer theory, Lancaster (1966) and Gorman (1980) developed the idea that an agent’s preferences over consumption goods depend on their characteristics. In philosophy, von Wright (1963) studied the logic of preferences, still influencing current work (e.g., Liu 2010); and Pettit (1991) and de Jongh and Liu (2009) discussed the dependence of an agent’s preferences on properties of the options.

In our own previous work, we developed a model of how reasons, or motivationally salient properties, relate to preferences, and used this model to study preference change (Dietrich and List 2011, 2013a, 2013b). Osherson and Weinstein (2012) proposed a formal logic of preferences based on reasons. Unlike these earlier papers, the present paper (i) focuses on the explanation of choice, not preference, (ii) treats motivationally salient properties, not as exogenously given, but as endogenously determined by the choice context, and (iii) considers not only “intrinsic” properties of the options, but also properties related to the choice context.

Structure of the paper: In Section 2, we briefly introduce the classical theory of rational choice, our point of departure. In Section 3, we informally describe our frame- work, followed by a more formal exposition in Section 4. In Section 5, we characterize all choice functions that can be explained in a reason-based way. In Section 6, we discuss some applications. In Section 7, we turn to the prediction of choices in novel contexts.

2Similarities to our reason-based approach can be found in Rubinstein’s (2006) distinction between

“internal” and “external” reasons for choice, in Manzini, Mariotti, and Mandler’s use of properties in checklists (as discussed later), and in the notions of “attention” or “consideration sets”, as typically discussed in relation to options rather than properties (e.g., Masatliogluet al. 2012).

(5)

2 The classical theory of rational choice

2.1 The basics

We begin by reviewing the basics of classical rational choice theory. The central concept is that of an agent’schoice function. This assigns, to each choice context, the option(s) chosen by the agent in that context. The aim is to explain or “rationalize” a given choice function by ascribing to the agent a preference relation over the options. This

“rationalization” is successful if, in each choice context, the agent chooses the most preferred option(s) in that context, according to the given preference relation.

It is natural to view the choice function as theexplanandum – the observable object that we seek to explain – and the preference relation as theexplanans – the theoretical object that does the explaining. However, as noted in the introduction, many choice theorists avoid using the language of “explanation”, because they interpret the pref- erence relation behaviouristically, as a mere representation of the choice function: a convenient way to express its informational content. Elsewhere, we have argued against this behaviouristic interpretation (Dietrich and List 2016).

Formally, the observable primitives of the classical theory are the following:

• A non-empty setX of options. Typical elements arex,y, z, ...

• A non-empty set K of contexts (sometimes called “menus”), where each element K 2 K is a non-empty setK ✓X of feasible options. In the simplest case, K is the set of all non-empty subsets ofX.

• Achoice function C:K!2X, which assigns to each contextK2Ka non-empty set of “chosen options” in K (i.e., C(K)✓ K). If the chosen set C(K) contains more than one option, this means that several options are tied for choice.

The choice function C is rationalizable by a preference relation if there exists a binary relation%onX such that, for all contextsK 2K,

C(K) ={x2K :x%yfor ally2K}.

A simple example illustrates these definitions. Here, the setXconsists of an apple, a banana, and a coconut; the setKconsists of all non-empty subsets ofX; and the choice function C is as follows:

• C({apple, banana, coconut}) ={apple};

• C({apple, banana}) ={apple};

(6)

• C({apple, coconut}) ={apple};

• C({banana, coconut}) ={banana};

• C({apple}) ={apple};

• C({banana}) ={banana};

• C({coconut}) ={coconut}.

This choice function can be rationalized by a (complete and transitive) preference relation

%which satisfies

apple banana coconut.

As is standard, is the strict part of%, and ⇠is the indi↵erence part.

2.2 When is a choice function rationalizable by a preference relation?

Not all logically possible choice functions can be rationalized by a preference relation.

For instance, if an agent chooses an apple from the set{apple, banana, coconut}and a banana from the set{apple, banana}, then no preference relation will rationalize this pat- tern of choices. To be consistent with the first choice, i.e.,C({apple, banana, coconut}) = {apple}, the preference relation would have to rank the apple at least weakly above all three fruits. But then the apple would also have to be chosen from the set{apple, banana}, which contradicts the second choice, i.e., C({apple, banana}) ={banana}.

From the perspective of scientific method, the fact that not all choice functions can be rationalized by a preference relation is good news. It means that the hypothesis that an agent’s choices are based on a preference relation is falsifiable; it is not a tautology (at least once the set of options has been fixed). The following classic result gives necessary and sufficient conditions for a choice function to be rationalizable by a preference relation.

Proposition 1 (Richter 1971) A choice function C is rationalizable by a preference relation if and only if it satisfies the axiom of Revelation Coherence.

To state that axiom, let us say that an option xis chosen weakly over an optiony in contextK ifx, y 2K and x2C(K). Further, xis chosen strictly overy inK if, in addition,y /2C(K).

Revelation Coherence For all contextsK 2Kand any feasible optionx2K, if, for every optiony2K, there is a context K02Kin whichxis chosen weakly over y, then x2C(K).

(7)

Revelation Coherence does not guarantee that the binary relation that rationalizes a given choice function satisfies any further properties such as acyclicity or transitivity.

For that, the choice function must satisfy stronger conditions, such as theWeak Axiom of Revealed Preference (e.g., Samuelson 1948; Bossert and Suzumura 2010). The details need not concern us here. What matters for our purposes is a general point: if, and only if, a choice function satisfies certain structural conditions, it can be rationalized by a preference relation.

2.3 Bounded versus sophisticated rationality

There are at least two familiar kinds of choice behaviours which conflict with the struc- tural conditions just mentioned and which the classical theory therefore cannot accom- modate – at least not without significant adjustments.

Cases of bounded rationality: As is empirically well established, human decision- makers often violate conditions such as Revelation Coherence or the Weak Axiom of Revealed Preference due to framing e↵ects, menu-dependent choice, susceptibility to nudges, the use of heuristics, unawareness, and other psychological phenomena. For example, a mere redescription of the options can lead to choice reversals. In Tversky and Kahneman’s framing experiments (e.g., 1981), participants reversed their choices over the same pair of options when their description was slightly modified, even though the experimenters were careful not to change any information conveyed. Similarly, policy makers are well aware that subtle changes in the decision environment, such as a change from an “opt-out” to an “opt-in” default in an insurance scheme, can greatly a↵ect people’s choices (Thaler and Sunstein 2008). Decision-makers also often satisfice rather than optimize or use simple heuristics (Gigerenzeret al. 2000). An example is someone whose rule of thumb for buying a banana is to choose one whose size is above the average of the batch on o↵er. None of these choices can be rationalized by a preference relation over the options, unless we redescribe the options in a complicated way.

Cases of sophisticated rationality: The structural conditions of the classical theory also fail to accommodate some intuitively rational but sophisticated forms of choice, such as choices based on norm-following or non-consequentialism. For example, a dinner-party guest who never chooses the largest piece of cake o↵ered to him or her for politeness and instead chooses the second largest cannot be rationalized by a preference relation over pieces of cake (Sen 1993). The classical theory deems this choice behaviour “irrational”, on a par with an ordinary rationality violation. Similarly, consider a professor who votes

(8)

for a university reform when the dean and president have respected the relevant proce- dures in the run-up to the vote, but votes against it when there has been a procedural breach. Assume that the reform and its consequences would be the same in both cases.

If the options are “reform” and “no reform”, we cannot rationalize this choice behaviour by a preference relation. To accommodate it, we would, at least, have to “re-individuate”

the options by building some features of the choice context into them.

We suggest that the classical theory’s difficulty in handling these cases, and its in- ability to distinguish bounded from sophisticated rationality, stems from the lack of a model of how agents perceive the options in any given choice context. When we provide such a model, a unified explanation of many of the challenging phenomena can be given.

3 Our framework, informally explained

3.1 The idea of a reason-based explanation of choice

Our basic idea is the following. When an agent chooses between several options in some context, e.g., yoghurts in a supermarket, he or she perceives each option not as a primitive object, but as a bundle of properties. Although each option can have many properties, the agent considers not all of them, but only a subset: the motivationally salient properties. In the supermarket, these may include whether the yoghurt is fruit- flavoured, low-fat, and free from artificial sweeteners, but exclude whether the yoghurt has an odd (as opposed to even) number of letters on its label (an irrelevant property) and whether it has been sustainably produced (a property ignored by many consumers).

The agent then makes his or her choice on the basis of afundamental preference relation over property bundles. He chooses one option over another in the given context, e.g., a low-fat cherry yoghurt over a full-fat, sugar-free vanilla yoghurt, if and only if his fundamental preference relation ranks the set of motivationally salient properties of the first option, say {low-fat, fruit-flavoured}, above the set of the second, say {full-fat, vanilla-flavoured, artificially sweetened}.

We call an agent’s choice behaviourreason-based explicable if it can be explained in this way. More precisely, areason-based explanation attributes two things to an agent:

• amotivational salience function, which assigns to each choice context the properties the agent cares about in that context: the “motivationally salient” properties; and

• afundamental preference relation over bundles of properties.

(9)

We call the pair consisting of a motivational salience function and a fundamental pref- erence relation areasons structure. According to a reason-based explanation, the agent perceives the options in each context through the lens of the motivationally salient properties in that context; and the agent then chooses an option whose bundle of moti- vationally salient properties he or she most prefers.

Later, we axiomatically characterize all choice functions that admit a reason-based explanation. Technically, reason-based explanation is a new rationalization concept.

But given our emphasis on the idea ofexplaining choices, we use the term “explanation”

rather than “rationalization”.

3.2 How the context matters

In our framework, the motivationally salient properties that occur in a reasons structure may be of up to three kinds:

• option properties, which options have independently of the choice context and which are thus “intrinsic” to the options;

• relational properties, which options have relative to the context; and

• context properties, which are properties of the context alone.

Examples of option properties are “fruit-flavoured” and “low-fat” (in yoghurts); these depend solely on the yoghurt itself. Examples of relational properties are whether a yoghurt is the only cherry yoghurt on display, or the cheapest; these depend also on the other available yoghurts. Examples of context properties are whether the available yoghurts include premium brands (this depends only on the menu) and whether there is background music (this depends on features of the context over and above the menu).

Reason-based explanations can capture two kinds of context-dependent motivation:

Context-variance: Here, the context a↵ects which properties are motivationally salient, so that the agent cares about di↵erent properties in di↵erent contexts. For example, some contexts make the agent diet-conscious, others not.

Context-relatedness: Here, the motivationally salient properties in some contexts go beyond option properties and include relational or context properties, so that the agent cares about the context or about how the options relate to it. For example, the agent cares about whether the choice of an option is polite in the given context, whether it is bigger than average, or whether there are luxury options available.

(10)

Many non-classical forms of choice can be subsumed under these two kinds of context- dependence. Arguably, bounded rationality, including susceptibility to framing, often involves context-variant motivation. Sophisticated rationality, such as norm-following or non-consequentialism, often involves context-related motivation. By contrast, classical rationality excludes both kinds of context-dependence. Of course, we do not claim that context-variance is always boundedly rational or that context-relatedness is always sophisticated. Our point is that reason-based explanations can be given for a variety of choice behaviours that are not classically rationalizable by a preference relation.

3.3 A common objection

Before we present our framework formally, it is worth addressing one common objection.

Since we take agents to perceive options as bundles of motivationally salient properties, a critic might ask why we do not simplydefine each option as a bundle of motivationally salient properties. Should we not define the set X as the set of all such bundles? A choice context would then be a set of property bundles among which the agent can choose. Everything else would remain classical.

There are, however, three problems with this proposal (see Bhattacharyyaet al. 2011 for some similar observations):

• First, we, the modellers, do not know in advance how the agent will perceive each option in a given context. The motivationally salient properties can be inferred, at most, after observing the agent’s choice behaviour.

• Second, an agent may perceive the same option through the lens of di↵erent prop- erties in di↵erent contexts, for instance when certain properties are motivationally salient in some contexts but not in others. This problem, together with the first, illustrates that, while we may treat options as observable primitives, we cannot equally treat an option’s motivationally salient properties as an observable prim- itive. The notion of motivational salience is invoked in our explanation of the agent’s choices; it is not part of our pre-theoreticdescription of those choices.

• Third, the same option can have di↵erent properties in di↵erent contexts when these properties are relational. For instance, the same piece of cake can be the second-largest in one context and the largest in another, and thus “politely choos- able” in the former context, but not in the latter. If we were to speak of two distinct pieces of cake here, we would no longer capture the fact that there is a perfectly intelligible sense in which they are the same, albeit in di↵erent contexts.

(11)

To address these problems, we must have a way of distinguishing between an option in the “objective” sense, as viewed from the “Olympian” perspective of the modeller, and an option in the “subjective” sense, as perceived by the agent whose choice behaviour we seek to explain. Our framework allows us to draw this distinction. We can think of each element of the original set X as an option in the “objective” sense. And we can think of each option’s bundle of motivationally salient properties in a given context as the option in the “subjective” sense, as perceived by the agent.

4 Our framework, formally defined

4.1 Observable primitives

We are now in a position to present our framework formally. The observable primitives are as in the classical theory. We have a non-empty setX of options; a non-empty set K of contexts, each of which o↵ers a non-empty set of feasible options (a subset of X);

and a choice function C : K!2X, which assigns to each context K 2K a non-empty set of chosen options among the feasible ones inK.

We permit only one small (but optional) generalization. Readers who do not like this generalization may ignore it; all our results also hold without it. We no longer require that each context beidentified with its set of feasible options. Instead, we merely require that it induce a set of feasible options. Thus a context K 2 K need not be a subset K ✓ X; it must merely pick out such a subset. This permits (but of course does not require) the existence of distinct contexts that o↵er the same options.

Specifically, each context K could be a pair (Y, ), whereY is the feasible set (with Y ✓ X) and is a parameter that specifies some further features of the environment (as in the notion of a “frame” or “ancillary condition” in Salant and Rubinstein 2008 and Bernheim and Rangel 2009; see Section 6.6 below). This parameter could represent a cue given to the agent, a specification of a “default” option, some priming before the choice, the cultural environment, some background music, or the room temperature – even a state of the agent such as “sober” or “drunk”. We might distinguish, for instance, between a supermarket with classical music in the background and the same supermarket with pop music, where there is no di↵erence in the goods on o↵er.

Officially, we write K for the context under our general definition, and [K] for its feasible set, so that [K] is a subset ofX, whileK need not be. For convenience, we often drop the square brackets and writeK for [K], since it is usually unambiguous whether K refers to the context itself or to the feasible set (e.g., in “x2K”,K refers to [K]).

(12)

4.2 Properties

Our next step is to define properties. At first, we might be tempted to define a property simply as a feature that an option may or may not have. Each property then picks out a subset ofX consisting of those options that have the property. The property “being a fat-free yoghurt” can be modelled like this. If X is the set of all possible goods in a supermarket, this property can be identified with the subset of X consisting of all fat- free yoghurts. However, this definition of properties is insufficiently general. As already noted, we want to allow for the possibility that an agent’s choices may be driven by properties that relate to the choice context.

We therefore define properties as features of option-context pairs, i.e., as features of pairs of the form (x, K), where xis an option and K is a choice context. Formally, a property is an abstract object, P, that picks out a subset [P] ✓ X ⇥K called its extension, consisting of all option-context pairs that “have” or “satisfy” the property;

thus properties are binary here. (X⇥K is the set of all option-context pairs.3) For convenience, we rule out properties that are never satisfied (i.e., [P] is the empty set?) and properties that are always satisfied (i.e., [P] is the universal setX⇥K).

Our definition allows distinct properties to have the same extension. This is use- ful for capturing framing e↵ects in which the description of a property matters. For example, the properties “80% fat-free” and “20% fat” (in foods) have the same exten- sion but di↵erent descriptions and may sometimes prompt di↵erent responses. In many applications, however, it suffices to identify properties with their extensions.

We can now formalize the distinction between option properties, context properties, and relational properties.

Option properties: These are properties whose possession by an option-context pair depends only on the option, not on the context; they are in this sense “intrinsic” to the option. Examples are “fat-free” and “vanilla-flavoured” (in yoghurts). Formally,P is an option property if

(x, K)2[P],(x, K0)2[P] for allx2X andK, K02K.

Context properties: These are properties whose possession by an option-context pair depends only on the context, not on the option. Examples are “o↵ering more than one feasible option”, “o↵ering a Rolls Royce among the feasible options”, and – if contexts specify the choice environment over and above the feasible set – the time (“it’s evening”),

3Some pairs (x, K) inXKare “infeasible” in the sense thatx /2K.

(13)

the temperature (“it’s a hot day”), or a default (“the status quo is such-and-such”).

Formally,P is a context property if

(x, K)2[P],(x0, K)2[P] for all x, x0 2X andK 2K.

Relational properties: These are properties whose possesion by an option-context pair depends on both the option and the context. Examples are “not being the largest piece of cake o↵ered” and “being the most expensive car on the market”. Formally, P is arelational property if it is neither an option property nor a context property.

We call properties that are not option propertiescontext-related and properties that are not context propertiesoption-related. Relational properties are context-relatedand option-related.

4.3 An example

To illustrate how properties can a↵ect choices, we give an example to which we will refer repeatedly. We introduce this example in a pre-theoretic way, and only later show how it can be explained in our framework. The example concerns the choice of fruit at a dinner party, as in Sen’s well-known story of a polite dinner-party guest (Sen 1993).

LetXcontain di↵erent fruits: apples, bananas, chocolate-covered pears, and possibly others. Each kind of fruit comes in up to three sizes: big, medium, and small. A choice context is a non-empty feasible set K ✓X, consisting of fruits currently in the basket (so, in this example, we require only the classical notion of a context). The set of possible contexts is K= 2X\{?}. We consider the following properties:

• “big”, “medium”, and “small”: the option properties of being a big, medium, and small fruit, respectively;

• “chocolate-o↵ering”: the context property of o↵ering at least one chocolate-covered fruit among the feasible options;

• “polite”: the relational property of not being the last available fruit of its kind, i.e., not being the last apple in the basket, the last banana, and so on.

We describe four agents whose choice behaviour we will later explain:

Bon-vivant Bonnie always chooses a largest available fruit. For anyK, she chooses C(K) ={x2K :xis largest in K},

where “medium” is larger than “small”, and “big” is larger than both other sizes.

(14)

Polite Pauline politely avoids choosing the last available fruit of its kind and only secondarily cares about a fruit’s size. For any K, she chooses

C(K) ={x2K :xis largest in K ifK6=? and largest inK ifK=?}, whereK is the set of all fruits inK that are not the last available ones of their kind.

Chocoholic Coco picks any fruit indi↵erently when no chocolate-covered fruit is avail- able, but otherwise chooses a largest available fruit, because the smell of chocolate makes him hungry. For anyK, he chooses

C(K) =

( K ifKcontains no chocolate-covered fruit, {x2K :xis largest in K} otherwise.

Weak-willed William makes the same polite choices as Pauline when no chocolate- covered fruit is available, and the same “greedy” choices as Bonnie otherwise, as the smell of chocolate makes him lose his inhibitions. For anyK, he chooses

C(K) = 8>

<

>:

{x2K:xis largest inK} if

"

Kcontains no chocolate-covered fruit andK6=?

# , {x2K:xis largest inK} otherwise,

whereKis again the set of fruits inKthat are not the last available ones of their kind.

4.4 Reason-based explanations

As already anticipated, a choice function admits a reason-based explanation if it can be explained by attributing a reasons structure to the agent. We now make this precise.

The set of potentially relevant properties: We begin by specifying a set P of potentially relevant properties. It contains the properties that we, the modellers, have at our disposal when we try to explain the agent’s choices. In our example,P might be the set {big, medium, small, chocolate-o↵ering, polite}. The set P can be partitioned into a set Poption of option properties, a set Pcontext of context properties, and a set Prelational of relational properties. Our specification ofPcan be viewed as a background hypothesis to the e↵ect that no properties outside P make a di↵erence to the agent’s choices, while at least some of the properties inside P might do so.4 Any subset ofP is

4Our criteria for specifying the setPmay be analogous to the criteria by which statisticians specify the potential explanatory and control variables in a regression analysis; i.e.,Pcan be specified permissively, but not unreasonably so. Defining P as the set of all logically possible properties, which contains a property for every proper subset ofXK, would not be good methodology, as explained later.

(15)

called a property bundle. For any optionxand any contextK, we further write

• P(x, K) for the bundle of all properties of (x, K), formally{P 2P: (x, K)2[P]};

• P(x) for the bundle of option properties ofx, formally P(x, K)\Poption; and

• P(K) for the bundle of context properties ofK, formally P(x, K)\Pcontext. A reasons structure: Areasons structure,R, is a pair (M, ) consisting of:

• Amotivational salience function M (formally a function from Kinto 2P), which assigns to each contextK 2Ka setM(K) ofmotivationally salient properties in contextK. We require the functionM to satisfy aninvariance constraint: if two contextsK andK0 are such thatP(K) =P(K0), thenM(K) =M(K0).

• A fundamental preference relation over property bundles (formally a binary relation on 2P, on which we initially impose no restrictions). We write > and⌘ for the strict and indi↵erence relations induced by .

The function M specifies which properties the agent cares about in each context, and the relation specifies how he or she cares about these properties, by ranking di↵erent property bundles relative to one another. The invariance constraint on M prevents an empirically ungrounded ascription of motivational di↵erences across contexts. It requires that any two contexts that have the same context properties induce the same motivationally salient properties. So, if we wish to hypothesize that the agent cares about di↵erent properties in contexts K and K0, we must be able to point to some di↵erence in context properties that lies behind this motivational di↵erence. Contexts that do not di↵er in their context properties should be motivationally indistinguishable.

How reasons explain choices: According to the reasons structure R= (M, ):

• The agentperceivesany optionxin any contextK as the bundle of motivationally salient properties of (x, K), denotedxK =P(x, K)\M(K).

• In any contextK, the agent willchoosethe options which, when perceived in terms of their motivationally salient properties in that context, are ranked most highly by his or her fundamental preference relation, formally

CR(K) ={x2K:xK yK for ally2K}.

We call CR (formally a function fromK into 2X) the choice function induced by R. If is insufficiently well-behaved,CR(K) may be empty for some K, so that CR may only be animproper choice function.

(16)

A choice function C : K! 2X is reason-based explicable if there there exists a reasons structureR(relative to the setP of properties) which induces that choice function (i.e., C =CR). We then callRareason-based explanation forC. Whether a choice function admits a reason-based explanation depends on the underlying setP of properties. We return to the significance of this dependence later.5

4.5 Revisiting the example

The four choice functions in our example all admit a reason-based explanation, where P= {big, medium, small, chocolate-o↵ering, polite}.

Bon-vivant Bonnie’s choice function can be explained by the reasons structure R= (M, ) where, for each context K,

M(K) ={big, medium, small}(soM is a constant function),

and the preference relation places the three singleton property bundles{big},{medium}, and{small}in the linear order satisfying

{big}>{medium}>{small}.6

For instance, in a contextKthat o↵ers only a small appleaand a big bananab, Bonnie perceives the two fruits as

aK =P(a, K)\M(K) ={small}, bK =P(b, K)\M(K) ={big}, and chooses the banana over the apple, because{big}>{small}.

5The agent’s fundamental preference relation over property bundles, which is context-independent, induces, for each contextK, a context-specific preference relation%Kover options: for anyxandyinX, x%Ky,xK yK. The choice functionCRcan therefore equivalently be defined as follows: for each K,CR(K) ={x2K:x%Ky for ally2K}.The equivalence betweenx%KyandxK yK is worth commenting on. In the expression “x%Ky”, options are understood “objectively” (as elements ofX), but the relation between them (%K) may depend on the context. In the expression “xK yK”, options are understood “subjectively” (as bundles of motivationally salient properties), but the relation between them ( ) is context-independent. The choice function induced byRcan thus be interpreted in two ways:

either as deriving from context-independent preferences over context-dependent (“subjective”) options, or as deriving from context-dependent preferences over context-independent (“objective”) options.

6Formally, = {({big},{big}), ({big},{medium}), ({big},{small}), ({medium},{medium}), ({medium},{small}), ({small},{small})}.

(17)

Polite Pauline’s choice function can be explained by the reasons structure R = (M, ) where, for each contextK,

M(K) ={big, medium, small, polite}(so, again,M is a constant function), and the preference relation places the property bundles{big, polite},{medium, polite}, {small, polite}, {big},{medium}and{small}in the linear order satisfying

{big, polite}>{medium, polite}>{small, polite}>{big}>{medium}>{small}. For instance, if only two small apples a and a0 and one big banana b are available in contextK, Pauline perceives the three fruits as

aK =P(a, K)\M(K) ={small, polite}, a0K =P(a0, K)\M(K) ={small, polite}, bK =P(b, K)\M(K) ={big},

and chooses one of the apples rather than the banana, because {small, polite}>{big}. Chocoholic Coco’s choice function can be explained by the reasons structureR= (M, ) where, for each contextK,

M(K) = 8>

>>

><

>>

>>

:

? if no chocolate-covered fruit is available inK, i.e., chocolate-o↵ering2/P(K),

{big, medium, if a chocolate-covered fruit is available inK, small} i.e., chocolate-o↵ering2P(K),

and the preference relation is the same as Bonnie’s, with the additional stipulation that?⌘?. For instance, in a context without a tempting chocolate-covered fruit, Coco picks any fruit indi↵erently, because he perceives every fruit as the same empty property bundle ?, where?⌘?.

Weak-willed William’s choice function can be explained by the reasons structure R= (M, ) where, for each context K,

M(K) = 8>

>>

><

>>

>>

:

{big, medium, if no chocolate-covered fruit is available inK, small, polite} i.e., chocolate-o↵ering2/P(K),

{big, medium, if a chocolate-covered fruit is available inK, small} i.e., chocolate-o↵ering2P(K),

(18)

and the preference relation is the same as Pauline’s. So, if context K o↵ers only two small applesaanda0and one big bananab, then, undisturbed by any smell of chocolate, William perceives these fruits as Pauline does and politely chooses a small apple. If a small chocolate-covered pear is added to the basket, he forgets about politeness and perceives the fruits as Bonnie does, choosing the big banana.

4.6 Two kinds of context-dependence

We say that an agent’s motivation, according to the reasons structureR= (M, ), is

• context-variant ifM is a non-constant function (i.e.,M(K) is not the same for all K2K), andcontext-invariant otherwise;

• context-related if the motivationally salient properties that are specified by M include context-related properties (i.e., M(K) contains at least one relational or context property for someK 2K), and context-unrelated otherwise.

In our example, Polite Pauline displays context-related motivation: the relational prop- erty “polite” is motivationally salient for her. Chocoholic Coco displays context-variant motivation: the properties that are motivationally salient for him vary with the con- text. Weak-willed William displays both kinds of context-dependent motivation: he sometimes cares about the relational property “polite”, and he also cares about di↵erent properties in di↵erent contexts. Bon-vivant Bonnie, finally, illustrates the classical case of fully context-independent motivation.

How do the two kinds of context-dependence a↵ect an agent’s perception of the options? Table 1 shows how a given option x is perceived in context K, depending on which of the two kinds of context-dependence are present. Generally, when both

Context-variant motivation?

Yes No

Context-related motivation?

Yes xK =P(x, K)\M(K) (e.g., William)

xK =P(x, K)\M (e.g., Pauline) No xK =P(x)\M(K)

(e.g., Coco)

xK =P(x)\M (e.g., Bonnie) Table 1: The agent’s perception of optionxin contextK

kinds of context-dependence may be present, option x is perceived in context K as xK = P(x, K)\M(K). This may depend on the context in two places: in the set of properties of the option-context pair (x, K) and in the set of motivationally salient

(19)

properties in contextK. If the agent’s motivation is context-unrelated, the first instance of context-dependence disappears, and P(x, K) can be replaced by P(x). Here,M(K) contains only option properties, so thatP(x, K)\M(K) =P(x)\M(K). If the agent’s motivation is context-invariant, the second instance of context-dependence disappears, andM(K) can be replaced by a fixed setM of motivationally salient properties. Here, the motivational salience function is constant, so that the first component of the reasons structure (M, ) can be identified with a fixed setM. In the context-independent case, finally, the agent’s perception of optionxin contextK simplifies to xK =P(x)\M.

From a classical perspective, agents with context-variant motivation – e.g., whose motivation varies as a result of subtle environmental features like the smell of chocolate – would count as boundedly rational. Bonnie exemplifies the case of classical rational- ity: her motivation is completely context-independent. Pauline displays sophisticated rational behaviour: she considers not only properties of the options, but also context- related properties, such as politeness. William tries to display the same sophisticated behaviour, but is susceptible to variations in motivation across di↵erent contexts. Coco, finally, focuses only on option properties, but, like William, lacks a stable motivation.

5 When does a choice function admit a reason-based explanation?

5.1 An axiomatic characterization

In what follows, we state three jointly necessary and sufficient conditions which a choice function C : K ! 2X must satisfy to admit a reason-based explanation. In line with convention, we call these conditions “axioms”, though we do not take their satisfaction for granted: it is an empirical question whether an agent’s choice function satisfies them.

Our axioms are each stated relative to a set P of properties. As already noted, whether there is a reason-based explanation for a given choice function depends on the set of properties we have at our disposal in constructing this explanation. Our axioms are jointly less restrictive ifPis rich than if it is sparse: it is easier to give a reason-based explanation if we have lots of properties at our disposal than if we have only a few.

We begin with an “intra-context” axiom. It says that the agent’s choice in any context does not distinguish between options with the same properties in that context:

Axiom 1 For all contexts K2Kand all optionsx, y 2K, ifP(x, K) =P(y, K), then x2C(K),y2C(K).

(20)

The second axiom is an “inter-context” axiom. It says that if two contexts o↵er the same feasible property bundles, the agent chooses options instantiating the same property bundles in those contexts:

Axiom 2 For all contextsK, K02K, if{P(x, K) :x2K}={P(x, K0) :x2K0}, then {P(x, K) :x2C(K)}={P(x, K0) :x2C(K0)}.7

Axioms 1 and 2 jointly imply that choice is based on the properties in P, but they do not yet imply any maximizing behaviour.8 This gap is filled by our third axiom, a variant of Richter’s original axiom of Revelation Coherence, as introduced in Section 2.

Unlike Richter’s axiom, ours is formulated at the level of property bundles, not options.

We adapt some revealed-preference terminology. For any property bundlesS andS0:

• S is feasible in contextK ifS=P(x, K) for some feasible optionx2K;

• S is chosen in context K ifS =P(x, K) for some option x2C(K);

• S is revealed weakly preferred to S0 (formallyS %C S0) if, in some context, S is chosen whileS0 is feasible.9

Axiom 3 Whenever a property bundle S ✓ P is feasible in a context K 2 K and is revealed weakly preferred to every feasible property bundle in context K, then S is chosen in contextK.10

Lemma 1 Axiom 3 strengthens Axiom 2.

We can now state our main characterization theorem:

7The axiom requires no relationship between choices in contexts with di↵erent context properties, i.e., whereP(K)6=P(K0), since such contexts automatically o↵er di↵erent feasible property bundles.

8They are jointly equivalent to choice being explicable by the attribution of a generalized reasons structure, defined by (i) a motivational salience function and (ii) a choice function defined over property bundles (which is more general than a fundamental preference relation over property bundles).

9The relation%C must not be interpreted as a fundamental preference relation. When the agent revealed-prefers bundleSto bundleS0by choosingSoverS0in some context, only somesubsetsofSand S0are usually motivationally salient, and the fundamental preference is held between these, not between SandS0. The revealed-preference relation%Cover property bundles induces a context-variant revealed- preference relation %CK over options, where x %CK y if and only if P(x, K) %C P(y, K). In classical choice theory, without properties, it is hard to define anycontext-variantrevealed preferences. Classical revealed preferences are context-invariant and fail to rationalize many observable choice behaviours.

10Like Axiom 2, this imposes “inter-context” contraints only among contexts with the same context properties: all contexts in which a given property bundle is feasible have the same context properties.

(21)

Theorem 1 A choice function C admits a reason-based explanation if and only if it satisfies Axioms 1 and 3 (and therefore 2).11

This result holds for every underlying set P of properties. We can thus use our framework to assess whether a given choice function admits a reason-based explanation relative to di↵erent sets of properties. We can ask: can we explain a car buyer’s choice function by reference to a set of colour-related properties? By reference to a set of status- related properties? Or by reference to a set of speed- and price-related properties? In each case, our axioms, relativized to the appropriateP, provide the required conditions.12 Reason-based explanations need not be unique. For a given choice functionC, there may exist more than one reasons structure Rsuch that C =CR. This non-uniqueness can be reduced if we impose further restrictions. In Appendix A, we state some additional characterization results, identifying conditions under which a choice function admits a reason-based explanation with only one, or none, of the two kinds of context-dependence we have discussed. Di↵erent reason-based explanations for the same choice function are by no means equivalent: they attribute a di↵erent motivational psychology to the agent and may lead to di↵erent predictions for novel choice contexts, as shown in Section 7.

5.2 The choice-behavioural falsifiability of reason-based explanations A key desideratum on any scientific theory is its falsifiability: it must be possible for the theory to be false. A theory that can “explain” everything does not explain any- thing. Theories of individual choice should be no exception. Choice theorists typically focus on choice-behavioural falsifiability. Although we think that there is no strong scientific reason to restrict the empirical evidence base to choice behaviour alone (ex- cluding, e.g., other psychological data), we temporarily follow convention and focus on choice-behavourial falsifiability too (cf. Dietrich and List 2016). How do reason-based explanations fare in this respect?

To answer this question, we must distinguish between two di↵erent senses in which reason-based explanations o↵er a theory of choice. On one interpretation, the specific reason-based explanation that we give for an agent’s choices is our theory. On another interpretation, thereason-based framework in its entirety is our theory.13

11Axioms 1 and 3 are jointly equivalent to the requirement that, for everyK2Kand everyx2K, if P(x, K) is revealed weakly preferred toP(y, K) for everyy2K, thenx2C(K).

12To make this explicit, we could restate Theorem 1 (and similarly other results) as follows: For every setP of properties, a choice function Cadmits a reason-based explanation relative to Pif and only if it satisfies Axioms 1 and 3 (and therefore 2) relative to P.

13A parallel distinction could be drawn in relation to classical rationalization concepts too.

(22)

A specific reason-based explanation as a theory: If an agent’s choice functionCis the observable object that we seek to explain, then the specific reasons structureRthat we attribute to the agent can be viewed as the theory that we o↵er as an explanation.

This theory, which we labelTR, has the form:

“R is the agent’s reasons structure (which impliesC=CR).”

This theory is clearly choice-behaviourally falsifiable. In particular, it is falsified if R fails to induceC (i.e., CR6=C) andcorroborated otherwise (i.e.,CR=C).

The reason-based framework as a theory: The broader message of our framework is that choices are “reason-based”. Applying this to a particular agent, we can view the assertion that the agent’s choice function C admits some reason-based explanation as our theory. Here, the theory, which we labelT9R,has the form:

“There is someR(relative to setP of properties) such that R is the agent’s reasons structure (which impliesC=CR).”

Whether this theory is choice-behaviourally falsifiable depends on the setPof properties relative to which it is asserted. If we are sufficiently disciplined in our specification of P, then T9R is choice-behaviourally falsifiable. With respect to many reasonable specifications of P (e.g., P = {big, medium, small, chocolate-o↵ering, polite} in our example), only some but not all choice functions satisfy our axioms for reason-based explicability. Hence T9R is falsified if the agent’s choice function violates our axioms, and corroborated otherwise. Note thatT9R is equivalent to the conjunction of Axioms 1, 2, and 3. By contrast, if we specify the setP too permissively, thenT9R may become choice-behaviourally unfalsifiable, as shown in the next subsection.

5.3 The significance of our auxiliary hypothesis

We have noted that the specification ofP is a crucial auxiliary hypothesis. It deems all properties that are outside that set irrelevant to the agent’s choices. This allows us to rule out reason-based explanations that are too far-fetched – for instance, because they invoke properties which do not plausibly matter psychologically, such as whether there is an even (rather than odd) number of letters on the yoghurt label. Far-fetched explanations, in turn, may not generate reliable predictions of future choices, as discussed later.

Let us illustrate how reason-based explicability will become too permissive and thereby substantively unilluminating if we specifyPtoo liberally. Suppose, for instance,

(23)

we takeP to include all properties of the form:

P(x,K): “The option-context pair is (x, K)”,

where x is an option andK a context in whichx is feasible. Let PX⇥K be the set of all such maximally specific properties – “maximally specific” because the extension of P(x,K) consists solely of the pair (x, K). It it easy to see that any logically possible choice functionC will admit a reason-based explanation whenever P ◆PX⇥K. Simply define R= (M, ) as follows:

• M(K) =PX⇥K for every context K;

• for any optionsxandyand any contextK, {P(x,K)} {P(y,K)}if and only ifxis weakly chosen overy in contextK.

In other words, Axioms 1 to 3 become vacuous whenP ◆PX⇥K, so that T9R becomes a tautology relative to such a setP.

However, the present reasons structureRdoes not provide an illuminating explana- tion of the choice function C. It accounts for the agent’s choices essentially by saying that the agent chooses option x over option y in context K because he or she funda- mentally prefers “xinK” to “yinK”. This is as unilluminating as saying “I preferred one to the other” when asked “why did you choose teaching rather than banking as your career”. A plausible auxiliary hypothesis would exclude maximally specific properties from the setP, unless we have special reasons to include them. Our goal is to identify properties that could make a psychologically plausible di↵erence to the agent’s choices.

5.4 Does the reliance on an auxiliary hypothesis make reason-based explanations ad hoc?

The reliance on an auxiliary hypothesis, encoded by P, does not render the notion of reason-based explanationad hoc. It is well known since the works of Duhem and Quine that practically all scientific theories rest on some auxiliary hypotheses. When we test a theory empirically, we are, in e↵ect, testing its conjunction with certain auxiliary hypotheses. Any apparently disconfirming evidence will seldom suffice to falsify the theory by itself, but will falsify it only relative to those auxiliary hypotheses. A stubborn supporter of the theory can always insist that the theory is correct and respond to the evidence by revising the auxiliary hypotheses.

This is famously illustrated by an episode from physics. In the 19th century, it be- came evident that Mercury’s orbit deviated from the one predicted by Newton’s theory.

(24)

But rather than admitting that Newton’s theory was falsified by this observation, some scholars, such as the mathematican Urbain Le Verrier, postulated the existence of an additional planet (“Vulcan”), whose gravitational influence would allow us to accommo- date Mercury’s orbit within Newton’s theory. Eventually, of course, Newton’s theory became overwhelmed with recalcitrant evidence, and it was superseded by Einstein’s.

Our claim is that the theory of individual choice is not di↵erent from other scientific theories in its reliance on auxiliary hypotheses. We have heard some people suggest (e.g., in response to this paper) that the classical notion of rationalization by a preference relation is purely choice-behavioural and free from auxiliary hypotheses. But this is not true. The key auxiliary hypothesis of the classical theory is its specification of the options. Although these are usually treated as exogenously given, the modeller implicitly asserts an auxiliary hypotheses when specifying them. Just as our notion of reason-based explicability becomes choice-behaviourally unfalsifiable when the setP of properties is specified too permissively, so the notion of rationalizability by a preference relation becomes unfalsifiable when the setX of options is specified too fine-grainedly.

To illustrate, let C (a function from K into 2X) be any choice function. Simply respecify the options as follows. LetX0 be the set of all pairs of the form (x, K), where x is an option in X and K is a context in which x is feasible. LetK0 be the result of replacing every original contextK inKwith

K0={(x, K) :x2K}.

Suppose we now reinterpret the original choice functionC as a functionC0fromK0 into 2X0 in the following way: for each K0 inK0, let

C0(K0) ={(y, K) :y2C(K), whereK is the context inKto whichK0 corresponds}. Then C0 will of course be rationalizable by a preference relation on X0, because each respecified option occurs in precisely one context. And this is so, whether or not the orginal choice function C was rationalizable by a preference relation on X. Crucially, from a choice-behavioural perspective, the functionsC andC0 are indistinguishable.

The upshot is this: by representing an agent’s choices in terms of a sufficiently fine- grained set of options, we can always “rationalize” any choice behaviour by a preference relation. And so, the hypothesis that the agent’s choices are rationalizable by a pref- erence relation is choice-behaviourally falsifiable only in conjunction with an auxiliary hypothesis, namely a hypothesis concerning the nature of the options. This issue is of- ten swept under the carpet. (For a notable exception, which includes a more elaborate formal argument for the point we have just made, see Bhattacharyya et al. 2011.) Our framework makes the role played by auxiliary hypotheses transparent.

(25)

5.5 Criteria for selecting an explanation in cases of non-uniqueness We have clarified the sense in which reason-based explanations are choice-behaviourally falsifiable. But we still need to comment on their possible non-uniqueness, relative to choice-behavioural evidence. How can we select a reason-based explanation when the same choice function can be explained in more than one way?14 This question matters because di↵erent explanations give di↵erent accounts of the agent’s motivational psychology, by attributing di↵erent reasons structures to him or her. These, in turn, may lead to di↵erent predictions for the agent’s future choices, as discussed in Section 7.

There are at least three kinds of criteria for deciding which reasons structureR= (M, ) to attribute to the agent when there are multiple competing ones:

Choice-behavioural di↵erence-making criteria: These require that, as far as possible:

(i) the motivational salience function M deem only those properties motivationally salient that make an observable di↵erence to the agent’s choice behaviour, and (ii) the fundamental preference relation over property bundles be systematically de-

rived from the agent’s choice behaviour.

The goal is to minimize behaviourally ungrounded ascriptions of motivation and funda- mental preference. We give one example of such a criterion in Appendix A.3.

Non-choice data: Verbal reports or neurophysiological data, such as responses to property-related stimuli, may help us test hypotheses about

(i) which properties are motivationally salient for the agent in context K and thus belong toM(K),

(ii) which context properties causally a↵ect motivational salience, so that M(K) may vary as contextsK vary in those properties, and

(iii) which property bundles the agent fundamentally prefers to which others.

One might hypothesize that human beings have better conscious access to how they perceive the options in a given contextK and therefore to the properties inM(K) than

14Non-uniqueness in the rationalization of choice behaviour is familiar from classical choice theory, where the same choice function can often be rationalized by more than one binary relation over the options. The relation becomes unique if the domain of the choice function (i.e., the set of contexts in which choice is observed) is “rich”, i.e., contains all sets of one or two options.

(26)

to the context properties that a↵ect what M(K) is (i.e., those properties which, in an empirical study, might be significant explanatory variables for M). Some changes in M(K) might be due to subconscious influences, as in framing or nudging e↵ects. If so, verbal reports may be more relevant to questions (i) and (iii) than to question (ii).

Parsimony criteria: We may try to select aparsimonious reasons structure, where (i) the setsM(K) of motivationally salient properties generated byM are (a) as small

as possible and (b) as unchanging as possible across di↵erent K, and

(ii) the relation is as sparse as possible (e.g., defined over the fewest possible property bundles).

There may be a trade-o↵ between di↵erent dimensions of parsimony. If the sets M(K) contain only few properties, they may not be stable across di↵erentK, and vice versa. As shown in Appendix B, we can alwaysformallyachieve context-invariance by definingM constantly as the entire setPand the fundamental preference relation as the revealed preference relation %C over property bundles. This makes the sets M(K) unchanging but very large, and hence perhaps psychologically implausible. Conversely, making each M(K) small might require context-variance.

5.6 Classical rationalizability as a special case

Finally, we wish to note that the notion of rationalizability by a preference relation can be recovered as a special case of reason-based explicability. Simply takeP=PX, defined as the set of all properties of the form

Px: “The option isx”,

where x is an element of X. The extension of each such property Px is the set of all option-context pairs in whichx is the option (i.e., [Px] ={(x, K) :K 2K}). Then the choice function C is classically rationalizable by a preference relation if and only if it can be explained by the reasons structureR= (M, ), where

• M(K) =PX for every contextK; and

• for any optionsxandyand any contextK,{Px} {Py}if and only ifxis weakly chosen overy in some contextK.

Of course, this explanation would be unilluminating, as it would always cite an option’s

“being that option” as the reason for choosing it. Nonetheless, the present observations help us compare the notion of reason-based explanation with its classical counterpart.

(27)

6 Some applications

To illustrate the generality of our framework, we briefly show how it can accommodate some much-discussed non-classical choice behaviours.

6.1 Framing e↵ects and choice reversals

As illustrated by Kahneman and Tversky’s influential work (e.g., 1981), a framing e↵ect occurs when an agent makes di↵erent choices in “extensionally equivalent” contexts, i.e., contexts which “objectively” o↵er the same options but which are somehow “framed”

(described, labelled, presented, ...) di↵erently. For instance, an agent may reverse his or her choice over public-health programmes, depending on whether these are framed in terms of the number of lives saved or the number of lives lost. Savingm out ofn lives (while not saving the remainingn m) is the same as losingn mout ofnlives (while saving the rest). Yet, people’s choice dispositions may depend on the wording used.

Formally, a framing e↵ect is a special kind of choice reversal. Achoice reversal occurs when there are contextsK andK0 and optionsxandysuch thatxis chosen overyinK andyis chosen overxinK0, where at least one choice is strict. Suppose R= (M, ) is the agent’s reasons structure in our framework. Then there may be two possible sources of choice reversals (as well as mixtures of the two).

• Context-variance: Here, the two contexts K and K0 in which a choice reversal occurs induce di↵erent sets of motivationally salient propertiesM(K) 6=M(K0), where bothM(K) andM(K0) contain only option properties.

• Context-relatedness: Here, contextsK andK0 induce the same set of motivation- ally salient properties M(K) = M(K0), but this set contains some relational or context properties that distinguish betweenxandyin the two contexts.

In either case, the agent prefers xtoyas perceived in context K, and prefersy toxas perceived in context K0, as illustrated in Figure 1.

Since framing e↵ects are usually thought to be subrational or subconscious, we may take a framing e↵ect to involve a choice reversal whose source is context-variance, not context-relatedness. Whether a choice reversal counts as a framing e↵ect so understood depends on the reasons structure we attribute to the agent. We may then define the frame in each contextK simply as the set of context properties ofK, formallyP(K).15

15If [K] = [K0], the di↵erence in frame can only be due to di↵erences in context beyond the feasible set, which presupposes our generalized (“non-extensional”) notion of context (as in Salant and Rubinstein

Références

Documents relatifs

This paper has three principal aims: to describe the development of a context-based course, Salters Advanced Chemistry, to draw together the research evidence on the effects of

This approach uses information from (i) acquisition technology (i.e., Microsoft’s Kinect) to adapt the choice of algorithms and their configuration to noised data, (ii)

Concrete, every value-based configuration unit, such as the necessary value in- terval of a supersonic sensor for a reconfiguration option like an autonomous parking service,

In this project, the most stringent scenario (from the control point of view) is the vision based landing of an aircraft on an unequipped runway whose size is partially unknown..

In the variability perspective of the SPES MF (ext.), the orthogonal variability model (OVM) can be used for modelling differences between product variants in terms of

The purpose of this paper is the research of the matrix-based knap- sack cipher in the context of additively homomorphic encryption and its appli- cation in secret

We then show how these newly defined measures can be used by agents to improve the quality of their understanding, and sketch a method in which agents can estimate them online

To be more precise, contributions of this paper are (i) the design, implementa- tion, and simulation of the inundation model on the wide area between the Hoa Binh dam and Hanoi